On the Rain and Drainage - Waters at Rothamsted. 
45 
In Table L. (p. 46) is given the estimated loss per acre of nitro- 
gen as nitrates in the drainage-water from the selected plots, for 
the two years commencing from the date of the spring sowing of 
the manures in 1879, to the corresponding period in 1881. 
Each year is divided into two periods, namely, from spring 
sowing to harvest, and from harvest to the next spring sowing. 
For each separate period, and for the whole period, the estimated 
total loss is given ; also the excess of loss from each plot receiving 
nitrogen in manure, over that from Plot 5 receiving mineral 
manure without nitrogen. The quantity of drainage for each 
period is assumed to be the same as that through the 60-inch 
soil-drain-gauge during the same period ; and the so assumed 
amounts of drainage are given at the bottom of the Table. 
The quantity of nitrogen as nitrates in the estimated drainage- 
water for each plot is reckoned according to the mean of all the 
analyses of the waters collected from the plot during periods 
from spring sowing to the end of May, from June 1 to harvest, 
from harvest to autumn sowing, and from autumn sowing to 
spring sowing. For some periods, however, for which there 
was drainage through the 60-inch soil-gauge, but none from the 
pipes in the experimental field, an estimate of the probable 
composition of the drainage shown by the drain-gauge has 
been made. Were it not that there is not always drainage 
from the drain-pipes when there is through the drain-gauge, 
it would be more correct to reckon each separate analysis on 
the amount of drainage from the 60-inch gauge. For this and 
other reasons, as will be seen further on, the estimates must be 
taken as only approximations to the truth. The nitrate of 
sodium on Plot 9, and the ammonium-salts (excepting on 
Plot 15), were sown, in 1879 on March 13, and in 1880 on 
March 9. On Plot 15, the ammonium-salts were applied on 
October 15, 1878 ; October 22, 1879 ; and October 25, 1880. 
Before referring to the estimates of the loss of nitrogen by 
drainage which the Table records, attention should be called to 
the fact that the two seasons in question were very exceptional 
in two ways, and do not therefore represent the average condi- 
tions of the experimental field. In the first place, in twenty-five 
of the first twenty-six of the thirty years during which (with a 
few special exceptions) the same description and amount of 
manure has been applied year after year on the same plot, the 
ammonium-salts were sown in the autumn ; but during the last 
four years, which include the two under consideration, they 
were (excepting on Plot 15) not applied until the spring. This 
fact would be supposed to tend to reduce the loss of nitrogen by 
drainage. Again, although the amounts of rain and drainage 
were very differently distributed in the two seasons, the 
